Annualizing a hot month is crypto’s version of planning the wedding after one good date. The conversation sparkles, the spreadsheet looks gorgeous, and suddenly you’re naming the children. Cute. Has anyone checked what happens when the attention leaves?
That was the joke in my quote of Ignas’s post about revenue-based token valuations. The serious bit deserves more room. Ignas described a new protocol whose apparently low valuation multiples depended on its last thirty days, while revenue was already falling. His question wasn’t whether on-chain revenue exists. It was whether the business behind that revenue survives. I’m not independently verifying his quoted multiples here, or identifying a token to buy. This is about the assumption hiding inside the denominator.
Annualization is a useful translation, not a time machine. Take revenue observed over a thirty-day window, divide it by thirty, then multiply the daily average by the number of days in a year. You now have a yearly run rate: what that pace would produce if it persisted. You do not have a year of observed revenue. You certainly do not have a promise about next year. The arithmetic can be perfectly correct while the story attached to it is doing cartwheels.
Why does that matter for a supposedly cheap token? A valuation-to-revenue multiple puts a valuation in the numerator and a revenue measure in the denominator. Inflate the denominator with an unusually busy period and the multiple gets smaller, without the token becoming cheaper in price. That can be a useful scenario. It becomes misleading when a screenshot quietly promotes the scenario into a durable earnings stream. My opinion: the more flattering the multiple, the more aggressively I want to interrogate the window.
Before interrogating the window, though, interrogate the label. DefiLlama defines fees as the total fees users pay to use a protocol. Its revenue measure is the subset retained by the protocol, excluding fees distributed to liquidity providers or other suppliers of capital. Revenue can go to a treasury, a team, or token holders. Those are not interchangeable destinations. DefiLlama explicitly describes its revenue definition as closer to gross profit in traditional accounting, not bottom-line company earnings. A chart labelled fees cannot simply be renamed token-holder income because the second label makes a better pitch.
Then follow the actual value path. Does a holder receive distributions, benefit from buybacks, or merely have governance over a treasury? What mechanism is live, and what is still a proposal? Who can change it? Protocol activity can be real while the token’s economic connection to that activity remains weak or conditional. Even a functioning buyback is not a guaranteed return to someone buying the token today. Purchase price, future supply and the durability of the underlying revenue still matter.
Now the boring months get their audition. I’d compare the recent window with longer available periods and inspect the daily series rather than just another annualized badge. Was activity spread across the month, or concentrated around a launch, a volatility spike or one popular trade? Were incentives encouraging people to show up? What changed in the fee schedule or product? These are research questions, not claims that any particular protocol manipulated its numbers. If a protocol is new, the longer history may simply not exist. Missing history is uncertainty, not an invitation to manufacture a backtest.
There’s another trap: calling gross revenue profit without looking at what sustains it. Incentives, security, development and operating commitments can matter even when they sit outside the displayed revenue metric. Token emissions may not leave a treasury as dollars, but that doesn’t make dilution irrelevant to holders. I’d also keep circulating market capitalization and fully diluted valuation clearly labelled rather than switching between them mid-comparison. Different numerators answer different questions; neither magically resolves future selling pressure.
None of this means every strong month is fake or every young protocol is doomed. A product can genuinely improve, win users and retain them. Annualized figures can help compare recent scale, provided the periods and definitions match. The bullish counterargument is that waiting for a long history can mean missing early growth. Fair. But accepting that trade-off means admitting you’re underwriting an uncertain future, not pretending the future has already arrived on-chain with an audit stamp.
My preference is a small set of explicit scenarios: recent activity persists, activity cools, or the product earns more repeat usage. I’d explain what evidence would support each, without assigning invented probabilities or dressing a spreadsheet up as a price target. Ignas’s useful point is that visible revenue doesn’t remove the need to judge a sector, a team and competition. Better data helps frame the bet; it doesn’t abolish it.
So when a token looks absurdly cheap on one hot month, I’m not reaching for a wedding planner. I’m asking what was measured, who gets paid, and what still works when the room gets quiet. That is a research standard, not a buy or sell recommendation. Show me the boring months. They tend to ask better questions.